The original application proposed a comprehensive characterization of morphogenesis in both wild type and experimentally manipulated embryos. Emphasis was to be placed on characterizing morphogenesis in both the notochord and the central nervous system. We also proposed to image experimentally manipulated embryos in which key embryogenesis genes were knocked down. In this revised two-year study we will restrict our efforts to the developing notochord in wild-type embryos. Our goal in focusing on the notochord is to establish the methodologies for image capture, segmentation and analysis that will be of broadest potential use to us, and others. In collecting the embryo images for this analysis we will be capturing the embryos in toto and in high resolution. Thus, while we will be focusing on the notochord, the image sets will be amenable to the analysis of other tissues in future studies. The notochord will serve as an ideal model for developing these methods. Besides the universal importance of the notochord for chordate development, the ascidian notochord presents a range of challenges for computer segmentation and analysis, yet it is probably the most tractable tissues in the embryo because of the low and fixed numbers of cells, and the fact that once cell intercalation is completed, the overall arrangement of the cells does not change. In the Preliminary Results section we demonstrated 3D time-lapse imaging of live embryos by both DIC and fluorescence microscopy. In this revised proposal we will continue with both of these imaging methods. As in the original proposal, several approaches for notochord cell segmentation will be pursued and optimized. We have presented preliminary results with two very different approaches for cell boundary detection: network snakes for DIC images;and a watershed-based method for confocal images of live embryos stained with a fluorescent membrane dye. We expect that concentrating on the notochord will allow us to make substantial contributions in only two years while building tools that can later be extended to more challenging tissues. Finally, we will develop tools for the morphometric analysis and visualization of the notochord data.